Alzheimer's disease (AD) is an age-related progressive disorder that affects 5 million people in the US. It is typically characterized by the presence of specific abnormal proteins in brain tissue that progress in a stereotypic fashion through different regions of the brain, starting in the hippocampus and spreading to the temporoparietal cortex. However, in some patients the progression of abnormal proteins does not conform to the typical pattern. There are cases which show very little involvement of the hippocampus (Hippocampal Sparing AD, HpSp AD), and those in which the proteins do not spread outside the hippocampus (limbic predominant AD). Little is known about these atypical variants yet they may account for a relatively large proportion of all AD cases (27%) and could influence the results of clinical studies on AD. The goal of this research project is to investigate how patients with atypical AD differ from those with typical AD. Aim 1 will assess clinical features, such as clinical diagnosis and performance on tests of cognitive functions such as memory and language, across the groups. The presence of specific genetic risk factors will also be assessed. Aim 2 will assess the patterns of brain tissue loss, i.e. atrophy, and how the brain shrinks over time across the groups using sophisticated automated techniques that analyze brain magnetic resonance imaging (MRI) scans. We expect that the patterns of brain tissue loss will differ across the groups, and will reflect the regions of the brain that the abnormal proteins are found at autopsy. In addition, aim 3 will use data from aim 1 and aim 2 and novel statistical techniques to determine whether this data could be used to predict the presence of atypical AD during life in an individual patient. This analysis will also serve to validate the distinction of atypical AD from typical AD. The grant will study 236 patients that have died and had a diagnosis of AD at autopsy. All patients will have been followed during life with yearly clinical assessments and MRI scans. The patients will be divided into three groups (HpSp AD, limbic AD and typical AD) based on the distribution of abnormal proteins in their brains at autopsy. Data concerning performance on specific clinical tests and presence of genetic biomarkers will be abstracted from the medical records and a clinical diagnosis will be determined by a behavioral neurologist. All MRI scans will be processed using three different types of analysis: patterns of brain tissue loss will be assessed using a technique called voxel-based morphometry;rates of change in brain volume will be assessed using image registration and the boundary-shift integral;and rates of change of particular structures in the brain will be assessed using Freesurfer software. Statistical analysis will be performed to compare results across groups for aims 1 and 2 and to perform cluster analysis for aim 3. The results from this project will provide data that will help clinicians to identify patients that may have atypical AD and may in the future lead to improved treatments for these patients. Since a high proportion of patients diagnosed with AD may have atypical AD this could lead to significant health improvements for many patients. PUBLIC HEALTH RELEVANCE: This study will characterize the clinical, genetic and imaging features of important, yet under recognized, atypical variants of Alzheimer's disease (AD) which will help clinicians identify these patients and will lead to improved treatments for such patients. Atypical AD is not uncommon, estimated to affect over 1 million Americans, and hence this study will have a significant impact on public health. Furthermore, identification of these atypical variants will likely have a positive effect on the outcome of clinical treatment trials which are designed to study typical AD patients.